What I am getting out of this conversation is that there is no point in doing anything sophisticated with these devices -- they simply are too unreliable. Is that the case? How can Silonex stay in business selling this item if it is so bad, and what applications are they being used for?
In a volume control role, they will mostly sit in the middle of their ranges, which appears to mean currents of a few milliamps or less, unless you turn the volume all the way down and leave it there for extended periods.
From the Silonex spec sheet my expectation was that you could light the LED to 20ma safely and leave it there indefinitely without damage, and only currents of more than 20ma would have deleterious effects on the device.
Is this unreliability a recent development caused by the change in manufacturing location mentioned a few posts back?
I wouldnt get to concerned about it. George hasnt had any failures. I have sold hundreds of quads. Heck, we are into the thousands of LDRs sold and havent had people complaining of them dying except the people who soldered them forever.
My only problem with them is the inconsistency in resistance which is due to them not using lasers to trim the resistive material or due to them using LEDs that are not lighting all exactly the same.
I think if I use a higher voltage and a lower current it will be very reliable as a switcher. They use far less than a milliamp at almost all volume settings as they only suck current at very low resistances, such as 250 or less. .5mA at 11V is about 250R so its easy to use them at low resistance without stressing the LED. I think they will live a long life like that.
Uriah
My only problem with them is the inconsistency in resistance which is due to them not using lasers to trim the resistive material or due to them using LEDs that are not lighting all exactly the same.
I think if I use a higher voltage and a lower current it will be very reliable as a switcher. They use far less than a milliamp at almost all volume settings as they only suck current at very low resistances, such as 250 or less. .5mA at 11V is about 250R so its easy to use them at low resistance without stressing the LED. I think they will live a long life like that.
Uriah
If you look at a PerkinElmer datasheet for the same sort of device, you can see the voltage vs current graph for a typical device. The voltage range is very narrow, and the current slope within that narrow voltage range is very steep. I think you need a very wide range of resistance to control the device across the very wide range of current specified for the 40~megohm resistances. At 5 volts supply and 20ma through the device to generate 40 ohms, you need 250 total system ohms (sum of current limit resistor plus LDR resistance), and at 5 volts and .1ma you need 50K ohms. I think the LDRs need control down to .01ma, so a 500K pot is possibly useful, but probably something in between.
If you change the voltage, you'll change the required resistance range. But the wide range of required resistance explains why the large value pot is required. You can lower the supply voltage by some amount and that will allow you to use a smaller value pot but you need to keep the current limiting resistor for when you turn the pot to minimum (zero) resistance. I think 100 ohms is not quite right for a 5V supply -- the correct resistor is going to be somewhere on the order of 120~130 ohms. At 100 ohms, you could be delivering 25ma or more to the LED, and that may explain the failures, I suppose.
If, for reliability, one must use a self-imposed current limit well below the specified 20ma and 40 ohms, then it appears to me that Silonex begins to lose appeal and the Perkin Elmer devices become competitive. The Perkin Elmer specifications certainly seem more believable -- detailed and consistent and well-explained. The output resistance vs input current is defined much more completely including reference to the effect of light memory on the relationship and they show multiple curves to demonstrate those differences. Further, they have a wide range of devices with different characteristics to choose from.
The VTL5C10 has a rated "on" resistance of about 75 ohms at 20ma which, if that device can be trusted at that current and the Silonex cannot, seems to be a better choice.
The "T" attenuator design I was contemplating uses three resistors. With a minimum device resistance of 40 ohms and Zo of 5K, that can provide an attenuation range of .14dB to 48dB, and at 75 ohms that becomes .3dB to 42dB, so the VTL5C10 is in the same ballpark as the Silonex device, if not quite as desirable.
The PerkinElmer unit is physically quite a bit larger than the Silonex, which could mean nothing, or it could mean a larger CdS surface for more consistent performance.
The PerkinElmer Application Notes are informative. There are graphs showing how distortion rises as voltage across the cell rises (and confirms that distortion is mainly 2nd harmonic), and recommends that average voltage across the resistive section be kept to 1 volt and below. A section on life and aging states that "output resistance will generally increase at a rate of 10 percent per year. The aging rate is lower with intermittent operation." I think that 10 percent is assuming max rated "on" for the year.
Geeze its like yelling fire in a theater.
Look I test thousands of these things a year.
My last test (June) of 460 of them ends with 12 that were thrown away. 7 of those were tossed out because they were so far out of tolerance that there is no use for them at all in any situation and 5 because at some point in testing they died.
I get pretty upset about this. This is about $20 in losses. We are talking about a 2.6% loss. For any device thats not acceptable. However it sure shouldnt have people running in the streets for fear of LDRs. Now, I still end up not using nearly half of them as they dont find a match. Then I leave them in my tester and fill it up again with new LDRs an d test again and end up finding a match for another 20-25% of the ones that didnt find a match in the first process. So, you can see my frustration as the lost dollars add up. What if you go out and buy 20 LDRs and match them? Will you find a good match? YES nearly without fail you will find a quad. Will you find two quads? Probably 50% of the time, yes. So is it worth it to try to test your own? Yes I think it is. Will you get a dead one out of 20? Probably not. Maybe. Who cares if you already got a quad? I dont think your quad will fail. Simply by the fact that they matched out of 20 total LDRs means that they are in the 'average' of the bunch and WERE manufactured properly and I would expect them to last a long time. Think of George's customers and their track record.
There is no other way that I know of to get sound this good in a passive attenuator. LDRs blow away Vishay Foil resistors, like in the Placette. LDRs blow away those TVC devices. LDRs blow away DACT and other super expensive stepped pots. Alps.. please.. they have no chance against LDRs.
Again, George has sold hundreds and no failures. None.. I have sold hundreds of quads, thousands of LDRs and no failures that I can attribute to anything but improper application of heat, and that being less than I can count on one hand.
I have had my original Lightspeed on for a few years now. I build a new LDR attenuator every month or two and for a year or so there I was building one every few days as I was trying different circuits. I am not scared one bit to put these in a commercial offering.
I tried Perkin Elmer. I dont get low enough resistance from them and they are 3, 4, 5, 6, 7 dollars each. To expensive to go matching. Silonex recently got an award for greatest contribution to music industry... I have to deal with their probably dated manufacturing process, but the good LDRs, which are most of them... work very well and sound better than any other passive in the world.
Does this scare a DIYer...? Seriously? Are commercial interests scared to offer LDRs from Silonex? NO. There are hundreds of products in the world in pro audio that have silonex LDRs in them. MOST OF THEM are not in an attenuator situation. I mean that they dont use two of them like we do. They use one because then they dont have to match anything. They still get great sound and if they got a bunch of returns dont you think Silonex would be out of business?
So here I am defending what I was upset with a few posts back. Simply because I was upset about small problems with them and it seems that you guys stretched that frustration into something like "OMG we should never use LDRs especially from Silonex!" which is absolutely wrong. I am so absolutely enamored with Silonex LDRs that I am absolutely sure I can send both of my daughters to college by designing circuits around them and selling them as quads for people who want to make their own attenuators.
In light of your comments though I will be more careful about how much current I give them. I right now in my own attenuator design give them only a max of 6mA each which gets me to my 40R safe place and if I use them in a switcher I will definitely be paralleling them so as to increase their life.
Basically, I think the response has just been a bit to over reactionary. Is that a word? Anyway, you get what I mean.
Uriah
Look I test thousands of these things a year.
My last test (June) of 460 of them ends with 12 that were thrown away. 7 of those were tossed out because they were so far out of tolerance that there is no use for them at all in any situation and 5 because at some point in testing they died.
I get pretty upset about this. This is about $20 in losses. We are talking about a 2.6% loss. For any device thats not acceptable. However it sure shouldnt have people running in the streets for fear of LDRs. Now, I still end up not using nearly half of them as they dont find a match. Then I leave them in my tester and fill it up again with new LDRs an d test again and end up finding a match for another 20-25% of the ones that didnt find a match in the first process. So, you can see my frustration as the lost dollars add up. What if you go out and buy 20 LDRs and match them? Will you find a good match? YES nearly without fail you will find a quad. Will you find two quads? Probably 50% of the time, yes. So is it worth it to try to test your own? Yes I think it is. Will you get a dead one out of 20? Probably not. Maybe. Who cares if you already got a quad? I dont think your quad will fail. Simply by the fact that they matched out of 20 total LDRs means that they are in the 'average' of the bunch and WERE manufactured properly and I would expect them to last a long time. Think of George's customers and their track record.
There is no other way that I know of to get sound this good in a passive attenuator. LDRs blow away Vishay Foil resistors, like in the Placette. LDRs blow away those TVC devices. LDRs blow away DACT and other super expensive stepped pots. Alps.. please.. they have no chance against LDRs.
Again, George has sold hundreds and no failures. None.. I have sold hundreds of quads, thousands of LDRs and no failures that I can attribute to anything but improper application of heat, and that being less than I can count on one hand.
I have had my original Lightspeed on for a few years now. I build a new LDR attenuator every month or two and for a year or so there I was building one every few days as I was trying different circuits. I am not scared one bit to put these in a commercial offering.
I tried Perkin Elmer. I dont get low enough resistance from them and they are 3, 4, 5, 6, 7 dollars each. To expensive to go matching. Silonex recently got an award for greatest contribution to music industry... I have to deal with their probably dated manufacturing process, but the good LDRs, which are most of them... work very well and sound better than any other passive in the world.
Does this scare a DIYer...? Seriously? Are commercial interests scared to offer LDRs from Silonex? NO. There are hundreds of products in the world in pro audio that have silonex LDRs in them. MOST OF THEM are not in an attenuator situation. I mean that they dont use two of them like we do. They use one because then they dont have to match anything. They still get great sound and if they got a bunch of returns dont you think Silonex would be out of business?
So here I am defending what I was upset with a few posts back. Simply because I was upset about small problems with them and it seems that you guys stretched that frustration into something like "OMG we should never use LDRs especially from Silonex!" which is absolutely wrong. I am so absolutely enamored with Silonex LDRs that I am absolutely sure I can send both of my daughters to college by designing circuits around them and selling them as quads for people who want to make their own attenuators.
In light of your comments though I will be more careful about how much current I give them. I right now in my own attenuator design give them only a max of 6mA each which gets me to my 40R safe place and if I use them in a switcher I will definitely be paralleling them so as to increase their life.
Basically, I think the response has just been a bit to over reactionary. Is that a word? Anyway, you get what I mean.
Uriah
OK!
Point taken, I won't take your gripes too seriously!On the other point about using 6ma to drive them to 40 ohms -- the Silonex chart suggests that 6ma is worth about 70 to 80 ohms, not 40 ohms. Is there that much variability? (I still haven't actually plugged one in yet, so I'm in no position to say from personal experience.)
I must be missing something in this explanation.
What controls the current at your desired 10mA?
Given the variability between individual samples of an LED, the only easy way to measure current flowing through an LED is to measure the voltage drop across a precision fixed resistor in series with the LED.
Because of the variability in LEDs, if you use a resistor that will limit a typical LED to 20ma, there will always be some LEDs that draw considerably more current due to lower than average internal resistance. That may be where your failures are coming from. The only way to avoid this problem is to use a resistor that will limit current to 20ma for most or all LEDs of a given type, which means that current through an average LED will be less than the optimal 20ma. You can also select a resistor value and measure the voltage drop across it to determine how much current a given LED is drawing at a given voltage. You could even "match" a resistor to a given LDR to insure 20ma at 5 volts.
At this point, admittedly not having applied power to any of my devices yet, I'd be willing to bet that 100 ohms is probably not enough current limiting for a 5V circuit. Even if the LED drops half of the 5 volts and the resistor only has to deal with 2.5 volts, that's still delivering 25ma across the circuit. If the LED drops only 2.0 volts, then the resistor is only limiting current to 30ma, which is way too much.
If you assume an LED drop of 2.0 volts out of a 5 volt circuit (as a worst case among LEDs with typical drops of 2.5V), you'll want to use a resistor of up to 150 ohms to limit current to 20ma for all LEDs.
Because of the variability in LEDs, if you use a resistor that will limit a typical LED to 20ma, there will always be some LEDs that draw considerably more current due to lower than average internal resistance. That may be where your failures are coming from. The only way to avoid this problem is to use a resistor that will limit current to 20ma for most or all LEDs of a given type, which means that current through an average LED will be less than the optimal 20ma. You can also select a resistor value and measure the voltage drop across it to determine how much current a given LED is drawing at a given voltage. You could even "match" a resistor to a given LDR to insure 20ma at 5 volts.
At this point, admittedly not having applied power to any of my devices yet, I'd be willing to bet that 100 ohms is probably not enough current limiting for a 5V circuit. Even if the LED drops half of the 5 volts and the resistor only has to deal with 2.5 volts, that's still delivering 25ma across the circuit. If the LED drops only 2.0 volts, then the resistor is only limiting current to 30ma, which is way too much.
If you assume an LED drop of 2.0 volts out of a 5 volt circuit (as a worst case among LEDs with typical drops of 2.5V), you'll want to use a resistor of up to 150 ohms to limit current to 20ma for all LEDs.
I'm looking at the impact of summing the taper of a volume control with the non-linear current vs resistance curve of the LDR itself, and it just looks like that causes a doubling of the taper.
Is anyone using a linear taper potentiometer to control these devices, and if yes, how is the perceived audio taper?
Is anyone using a linear taper potentiometer to control these devices, and if yes, how is the perceived audio taper?
OK!
On the other point about using 6ma to drive them to 40 ohms -- the Silonex chart suggests that 6ma is worth about 70 to 80 ohms, not 40 ohms. Is there that much variability? (I still haven't actually plugged one in yet, so I'm in no position to say from personal experience.)
Well, remember this.. I dont run at just 5V. The voltage matters to. Plug one in!!
vary the voltage and vary the current with a big 10k pot or maybe 20k. Its going to show you a lot. You might even decide if you like varying the current or the voltage better. Maybe graph it out for yourself.... voltage variance/resistance vs current variance/resistance. Dont worry with that data sheet from Silonex. Its not representative of very much at all. The resistance increases much faster than they suggest.ANDREW: I think he was using the LM317 as a current source.
Uriah
I must be missing something in this explanation.
What controls the current at your desired 10mA?
------
as mentioned, i used a 317 constant current source to set the current.
i hope i'm not spoiling anyone's fun. i reported my experience for a particular usage of these LDRs. it does not mean that they will have the same problem with other usage. this happened last year and i switched to something else since. i may be just unlucky with a bad batch. who knows?
What controls the current at your desired 10mA?
------
as mentioned, i used a 317 constant current source to set the current.
i hope i'm not spoiling anyone's fun. i reported my experience for a particular usage of these LDRs. it does not mean that they will have the same problem with other usage. this happened last year and i switched to something else since. i may be just unlucky with a bad batch. who knows?
It is clear and it is clearly wrong.
The Light Emitting Diode (LED) is fed with current and this determines the light output of the LED.
The light output shines on the Light Dependant Resistor (LDR).
The quantity of light determines the resistor value.
If the LED is not powered and you apply 10mA to the LDR you will burn it out. See spec sheet.
The Light Emitting Diode (LED) is fed with current and this determines the light output of the LED.
The light output shines on the Light Dependant Resistor (LDR).
The quantity of light determines the resistor value.
If the LED is not powered and you apply 10mA to the LDR you will burn it out. See spec sheet.